Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyze behavior across scales — detect regime changes, identify capacity limits, fit scaling laws within regimes.
.claude/skills/yogsoth-ai-scaling-frontier/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 45% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 108% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -57% | 0% |
Detect where scaling behavior qualitatively shifts.
| Base SOP | Target | ±10% Range | |----------|--------|------------| | web-search | 30 | 27–33 | | web-research | 10 | 9–11 | | paper-overview | 30 | 27–33 | | paper-search | 20 | 18–22 | | paper-research | 15 | 13–17 |
<HARD-GATE>
| SOP | Done | Target | % |
|-----|------|--------|---|
| web-search | ? | 30 | ? |
| web-research | ? | 10 | ? |
| paper-overview | ? | 30 | ? |
| paper-search | ? | 20 | ? |
| paper-research | ? | 15 | ? |
Budget Gate: OPEN/CLOSED (>=80% required to exit)
</HARD-GATE>Import: web-search, web-research, paper-overview, paper-search, paper-research Subagent: scaling-regime-detection, edge-case-generation
Analyze behavior across scales (data size, model size, compute, time), detect regime changes (breakpoints where behavior qualitatively shifts), identify capacity limits and their mechanisms.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | edge-case-generation | Systematically generate boundary inputs — boundary values, adversarial constructions, distribution shifts, rare combinations, scale extremes. | | scaling-regime-detection | Detect regime changes in scaling behavior — breakpoints where behavior qualitatively shifts, mechanisms behind transitions. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 42,171 | 78,390 | +86% | 1 | 1 | 0% | 6,019 | 8,714 | +45% | 0 | 0 | — |
case-02 | fail→pass | 26,469 | 66,205 | +150% | 1 | 1 | 0% | 4,180 | 8,696 | +108% | 0 | 0 | — |
case-03 | fail→pass | 54,722 | 52,018 | -5% | 1 | 1 | 0% | 8,269 | 8,707 | +5% | 0 | 0 | — |
case-04 | pass→fail | 34,145 | 46,420 | +36% | 1 | 1 | 0% | 5,026 | 8,678 | +73% | 0 | 0 | — |
case-05 | pass→fail | 20,220 | 43,055 | +113% | 1 | 1 | 0% | 3,453 | 7,509 | +117% | 0 | 0 | — |
case-06 | pass→pass | 28,663 | 40,586 | +42% | 1 | 1 | 0% | 4,673 | 7,820 | +67% | 0 | 0 | — |
case-07 | fail→pass | 23,530 | 20,638 | -12% | 1 | 1 | 0% | 3,307 | 3,209 | -3% | 0 | 0 | — |
case-08 | fail→pass | 39,979 | 7,128 | -82% | 1 | 1 | 0% | 1,628 | 696 | -57% | 0 | 0 | — |
case-09 | fail→pass | 31,920 | 7,774 | -76% | 1 | 1 | 0% | 1,729 | 878 | -49% | 0 | 0 | — |
case-10 | fail→pass | 17,705 | 12,318 | -30% | 1 | 1 | 0% | 2,098 | 1,734 | -17% | 0 | 0 | — |
case-11 | fail→pass | 62,314 | 50,804 | -18% | 1 | 1 | 0% | 1,321 | 8,648 | +555% | 0 | 0 | — |
case-12 | fail→pass | 45,110 | 2,840 | -94% | 1 | 1 | 0% | 932 | 888 | -5% | 0 | 0 | — |
case-13 | pass→pass | 18,210 | 4,011 | -78% | 1 | 1 | 0% | 1,961 | 1,129 | -42% | 0 | 0 | — |
case-14 | pass→pass | 13,962 | 8,155 | -42% | 1 | 1 | 0% | 1,869 | 1,035 | -45% | 0 | 0 | — |
case-15 | fail→pass | 78,698 | 47,990 | -39% | 1 | 1 | 0% | 5,975 | 8,678 | +45% | 0 | 0 | — |
case-16 | fail→fail | 79,310 | 44,439 | -44% | 1 | 1 | 0% | 8,230 | 8,668 | +5% | 0 | 0 | — |
case-17 | fail→pass | 49,620 | 54,072 | +9% | 1 | 1 | 0% | 8,237 | 8,675 | +5% | 0 | 0 | — |
case-18 | fail→fail | 34,644 | 51,821 | +50% | 1 | 1 | 0% | 5,347 | 8,681 | +62% | 0 | 0 | — |
case-19 | fail→pass | 52,313 | 44,262 | -15% | 1 | 1 | 0% | 8,239 | 8,677 | +5% | 0 | 0 | — |
case-20 | fail→pass | 32,379 | 56,524 | +75% | 1 | 1 | 0% | 5,645 | 8,675 | +54% | 0 | 0 | — |
case-21 | fail→pass | 42,551 | 62,974 | +48% | 1 | 1 | 0% | 6,091 | 8,668 | +42% | 0 | 0 | — |
case-22 | fail→pass | 33,571 | 44,476 | +32% | 1 | 1 | 0% | 5,353 | 8,673 | +62% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +59 percentage points is the difference between those two pass rates over the 19 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.